squall/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

[V1] Fix EngineArgs refactor on V1 (#9954)

Browse Source
This commit is contained in:
Robert Shaw 2024-11-02 10:44:38 -04:00 committed by GitHub
parent e893795443
commit d6459b4516
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 12 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
vllm/v1/executor/gpu_executor.py
View File

@ -1,10 +1,7 @@
import os import os
from typing import Optional, Tuple from typing import Optional, Tuple
from vllm.config import (CacheConfig, DeviceConfig, LoadConfig, LoRAConfig, from vllm.config import EngineConfig
ModelConfig, ObservabilityConfig, ParallelConfig,
PromptAdapterConfig, SchedulerConfig,
SpeculativeConfig)
from vllm.logger import init_logger from vllm.logger import init_logger
from vllm.utils import get_distributed_init_method, get_ip, get_open_port from vllm.utils import get_distributed_init_method, get_ip, get_open_port
from vllm.v1.outputs import ModelRunnerOutput from vllm.v1.outputs import ModelRunnerOutput
@ -15,29 +12,17 @@ logger = init_logger(__name__)
class GPUExecutor: class GPUExecutor:
def __init__( def __init__(self, vllm_config: EngineConfig) -> None:
self, self.model_config = vllm_config.model_config
model_config: ModelConfig, self.cache_config = vllm_config.cache_config
cache_config: CacheConfig, self.lora_config = vllm_config.lora_config
parallel_config: ParallelConfig, self.load_config = vllm_config.load_config
scheduler_config: SchedulerConfig, self.parallel_config = vllm_config.parallel_config
device_config: DeviceConfig, self.scheduler_config = vllm_config.scheduler_config
load_config: LoadConfig, self.device_config = vllm_config.device_config
lora_config: Optional[LoRAConfig], self.speculative_config = vllm_config.speculative_config
speculative_config: Optional[SpeculativeConfig], self.prompt_adapter_config = vllm_config.prompt_adapter_config
prompt_adapter_config: Optional[PromptAdapterConfig], self.observability_config = vllm_config.observability_config
observability_config: Optional[ObservabilityConfig],
) -> None:
self.model_config = model_config
self.cache_config = cache_config
self.lora_config = lora_config
self.load_config = load_config
self.parallel_config = parallel_config
self.scheduler_config = scheduler_config
self.device_config = device_config
self.speculative_config = speculative_config
self.prompt_adapter_config = prompt_adapter_config
self.observability_config = observability_config
self.worker = self._create_worker() self.worker = self._create_worker()
self.worker.initialize() self.worker.initialize()